Building something we can be proud of: the new FT office in Sofia
By Jon Furse
Our new Product & Technology organisation in Sofia, Bulgaria, is now 12 months old, with 75 staff and growing. We have nearly completed the transition of a large and operationally important part of our technology estate from London to the new teams in Sofia and we moved into our amazing new office in the centre of the city in February this year.
From the beginning, it was very important to me that our new offshore organisation be equal in status to our head office in London, have an equivalently high level of autonomy and for us to retain our talented London staff, who were to be highly involved in the transition. I think we have achieved extraordinary success in these endeavours, in what was a very challenging undertaking. This is the story of why and how we embarked on this project.
Finding The Right Strategy
In 2016, the Product & Technology department of the Financial Times (FT) were given three years to make a sizeable cut in capital expenditure, whilst continuing to target rapid business growth. Finding savings here and there wouldn’t amount to enough and would potentially jeopardise growth. We saw this as an opportunity to get on the front foot and do something different.
After carefully considering our options, we decided the boldest choice would be to invest in a new offshore office. Investing in a location outside of London was never going to be an easy option but one that had the biggest potential for future cost savings, since London is still one of the most expensive cities in the world to do business in. We were confident we wanted to retain a significant London presence for our Product & Technology department given our exceptional talent at the FT, extensive talent market and proximity to commercial and editorial leadership. But given the high costs, London was not the place to increase our efficiency to meet a cost challenge.
We also had a history of building offshore teams in our Manila office as well as working with partners in India and Romania but these relationships had grown organically with little strategic oversight and although our staff and partners were individually good, the teams were often too piecemeal and aligned to specific expertise. Investing out of London would therefore require a reorganisation.
In Product & Technology we are structured into groups, centred around customers and technologies. For instance FT Customer Products is responsible for FT.com and Internal Customer Products oversees the development of commercial and editorial tools. We considered whether it would work to transition any one of those groups, and concluded that it would be possible with FT Core. FT Core is the group that owns the technology platform that powers our customer-facing products, with capabilities spanning content publishing, content metadata, paywall and analytics data — and the group which I lead. FT Core was selected because of its existing relationship with offshore offices and partners, which we felt we could make more efficient, and because of its fragmented heritage, large scope, size and importance to the other groups.
Setting Clear Principles
It was important that our new offshore office should be more than just an engineering centre of excellence. It had to have a clear purpose for Product & Technology as a whole and we wanted the Sofia office to own the development of products and services, working with existing London and Manila teams across other groups to deliver against shared target outcomes. We created the following principles to guide us:
- One-team — our offshore office must be a part of the whole Product & Technology department. Employees in our offshore office must have equal status with colleagues in London and Manila
- Diverse and inclusive — we wanted to ensure consistency across our offices and support the FT’s drive towards creating a diverse and inclusive workforce, setting us apart from other offshore organisations
- Equality across disciplines — product, business analysis, engineering and delivery-lead roles must all be represented in our new office and they must all contribute to the broader success
- Equality in hierarchy — employ senior roles across all disciplines in the offshore office where possible to promote autonomy
- Propagate empowerment culture — empower our engineers and product managers in our new offshore office to select the ideas, tools and technologies with which to build technical solutions that meet our business outcomes
- Delivering value is king — our offshore office should increase the efficiency and productivity of delivery within Product & Technology
Selecting a Delivery Model
The FT has offices all over the world and our Manila office celebrated its 20th birthday in 2018, so setting up offices in new locations is something the FT has significant experience in. Yet a do-it-yourself solution had little appeal for establishing a new Product & Technology office — given the competitive global tech market, we felt strongly that we needed an existing local presence with strong relationships in the local market. We considered a range of options including acquisition, consultancy and a joint venture with a partner organisation.
We quickly threw out ideas of acquisition (too challenging to quickly adopt a new identity), consultancy (we felt strongly that this was something we needed to do in consultation and partnership with our staff, not something that would be done to them by untrusted external entities) and joint venture (this didn’t seem to make sense for FT Core which provides unbilled for services to other FT groups). One remaining model stood out — BOT.
In IT offshoring, Build Operate Transfer (or BOT) describes a relationship in which an organisation hires another organisation as a service provider to set up and run an offshore operation, with the intent of transferring the operation to the organisation as its own offshore facility in the near future. The benefit to the FT in this model would be an experienced partner with deep local knowledge of the market to navigate the initial build but with the ultimate benefit of bringing the entity back into full control of the FT at transfer time.
Selecting the Location
Initially all locations were on the table but we concentrated first on countries we already had a relationship with — India, the Philippines and Romania. We explored our own networks and visited key individuals and organisations in many countries to learn and discover what might work for us. Some of these organisations in turn came to visit us in London.
During the discovery phase, our engineers spent time with engineers from candidate partners, sharing architecture, processes, tools and structures. We set out to be as flawsome as possible — sharing both our great achievements and our worst failures so that no false expectations were set — and we made the sessions as friendly, open and adaptive as we could. With each iteration we refined our ideas and our wants and needs. It took a number of months but it was an incredibly enriching experience.
We found out quickly that wherever we went, engineering talent was already in great demand and that with greater distance from London it became cheaper but more complicated due to a variety of factors including time zone, travel distance, cultural differences and brand affinity. We came to recognise that proximity to London was important.
Using our research, we derived a score sheet of things that were important to us, including an experienced partner with similar culture to the FT and a low cost location with good infrastructure. After a long search and a couple of false starts, we found an amazing partner in Xoomworks, based in the Bulgarian capital city of Sofia. Sofia is a thriving city of over a million, with a burgeoning tech scene, a three hour flight from London airports and one of the best value locations in Europe. Xoomworks has a great heritage in similar projects, an established Sofia office and close cultural alignment with the FT. We agreed to start our contract negotiations with them as soon as possible.
The Business Case
My team in London are incredibly talented and we were clear upfront that we would work as hard as possible to retain them during and after the transition. So our business case focused on our contractor headcount and costs, which had steadily risen over a period of time, both in London and offshore. We felt strongly that by investing in our own staff (through BOT) in a close offshore location, whilst replacing London contractors with our own London staff across the remaining groups in London, we could both save money through cheaper per capita operating costs and increase productivity through consolidating relationships.
We built the business case with Xoomworks in around ten weeks. It involved deep and highly constructive collaboration between senior management teams in both organisations. During this period we worked with Xoomworks to build a small team in Sofia to establish the relationship and test our ways of working. In April 2018, we completed the business case, gained investment signoff and received board approval. After 18 months of research and preparation, we were ready to go.
Delivering the Change
We are extremely proud of our open, caring and inclusive culture. So being open with all our staff about the full scope and goals of this project was vital from the start and we worked hard on our strategy to communicate the full context and purpose of the change, including managing the impact to groups outside of FT Core.
I recognised early in the planning stage that what we were going to do was big. To make the change manageable, I phased the knowledge and technical transfer along existing teams, with each phase lasting 6–9 months, depending on complexity, and with a new phase starting every 3 months. This in turn defined the recruitment cycle and the way we worked with our existing partner organisations.
Over an estimated 15 months, we needed to transition the operations and delivery of more than 400 critical services and a broad range of tech from disparate teams summing 100 roles, to a new organisation and location where there were initially no staff. It was a huge ask of our staff to lead the transition of their considerable technical and domain knowledge for a lengthy period of time to an unknown team and we knew it would be hard for all FT Core staff and especially for those teams that were to transition last.
A key risk to the change was maintaining high service levels for core services, once all the existing engineering experience in London had moved onto new roles. To mitigate this risk as much as possible, we wanted the new teams to be able to demonstrate ownership and responsibility for their technical estate in Sofia. We defined a point, the “Assumption of Responsibility (AoR)” for each of the phases where our data would indicate the team had full autonomy and mastery of their part of FT Core. The FT would be satisfied that AoR had occurred when the team in Sofia was meeting a set of target metrics which correlated with ownership, including committing and reviewing a percentage of the codebase, managing the backlog and handling all incidents.
We built some processes to extract this information from our Git, Jira and Salesforce systems and began to report against these numbers as soon as we had meaningful data. The AoR data did provide useful insights — for example, by showing that London were still reviewing one team’s code, we were able to nudge the Sofia team to release their dependency on the London team and we were able to spot a potential quality issue related to a rising number of incidents.
To plan the technical transition, each team built and managed their own Kanban board of transition tasks. Each board had the same set of workflow states, including: high level overview, pairing between individuals and reverse knowledge transfer — where the receiving team played back their knowledge to the departing experts. Tasks could be any meaningful items, such as: high level overviews of a particular process, deep technical dives into specific systems or how to manage 24/7 support processes. We monitored these tasks using cumulative flow diagrams to ensure each phase was tracking on time and there were no bottlenecks.
Engagement between the new teams and departing teams was necessarily deep and sustained. Asking our London teams to transition their hard-won technical and domain knowledge to new staff at the same time as trying to secure a new role elsewhere in the FT was demanding. So we invested in bi-directional travel to allow connections to grow stronger between Sofia and London. We asked for volunteers to spend significant time in Sofia to help not just with the technical transition but also to seed our culture there. We created an incentive plan to encourage staff to continue the transition to the end of each phase and we promised to help train staff if required for their new role.
To help staff find the best possible new roles, we set up a sponsorship programme in London, with senior London staff taking time to help individuals find the most attractive group and team to transition into. By providing an opportunity to try one or more teams for up to two weeks’ trial and not requiring any formal selection procedure, the psychological impact of moving to a new team was lessened. As long as we could fill an existing vacancy or swap a London-based contractor, then we tried to be as flexible and accommodating as possible. We listened to feedback and where necessary, we adapted and compromised, always with the dual aim of securing both an autonomous Sofia office and retaining our talented staff in London.
2019 was the first year that the expected savings were due to be made and this year’s investment round saw the successful transition to Sofia make a significant difference to our ability to plan for our sought-after growth. Now, as we near the end of the transition, I have been reflecting on what we’ve achieved and what we’ve learned in the last 12 months.
- Bias for action — launching the programme in April was important for financial reasons and we made a ‘go’ decision knowing that we hadn’t dotted all the i’s and crossed all the t’s up front. Consequently we missed some items in the business case — like furniture in the new office! However, the bias for action served us well and we were usually able to adapt to missing requirements or uncertainties and our plans evolved during the year as we discovered more information
- Staff ownership — by not bringing in an external organisation to run the programme, we were explicitly asking our staff to own it. We felt sure our staff would provide more expert guidance, facilitate cultural knowledge transfer better and would be more engaged that way. But it was a big ask when they had so much invested. It was the right thing for the FT, where staff really feel strongly about the FT’s mission but it might not work with all organisations
- Staff support — providing support to all staff involved in the transition was vital. We invested early in groups to keep track of team sentiment, Q&A sessions and time from senior management. However, we should have invested earlier in work that would have provided more certainty for our London staff — like ready-to-roll incentive and sponsorship plans
- Onboarding — we initially struggled to efficiently onboard our staff in Sofia because we didn’t have the processes to handle the scale of new recruits (up to 10 a month into a small initial cohort). Facebook and many other companies have shown that investment in onboarding pays off in employee satisfaction and it’s an area we are now putting considerable effort into
- Communication impact — at launch we communicated our intentions to the entire business. We knew there would be a delivery impact whilst the teams transitioned and built up expertise but we probably could have done more to keep all stakeholders informed of status as the programme progressed. Likewise, expectations for problem resolution times for the new teams were initially unrealistic and sometimes required help from other teams, which we could have foreseen.
- Escalate early — remarkably, three months into our recruitment drive in Sofia, we’d managed to secure only three staff! Xoomworks sponsors escalated and we worked together on a way forward. Repeatedly holding each other to account and working closely on such challenges build a high degree of trust between our two organisations. Nine months on and we’ve build an enormously trusting partnership (and have more than 70 new staff!)
We’re near the end of the scheduled programme and I am feeling very positive about the change. The commitment and professionalism of our staff and contractors throughout has been incredibly inspiring and to me, it is clear that the transition would not have been possible without their deep and positive engagement. We lived by our principles and now have a diverse, highly talented and autonomous new team in Sofia. A new environment, bigger teams and clearer priorities are already providing dividends in the delivery of change in FTCore.
The project savings, primarily attributed to a reduction of our London contractor base, have contributed to a significant saving for 2019. Next year the forecast savings increase further to the planned 20% of our capital expenditure budget.
Transitioning FT Core to Sofia hasn’t just been about saving money though. It’s also provided opportunities to reevaluate almost everything — from the ways our teams work together, to the vision and mission of the FT Core group and the technical stacks we work with. It has enabled the business to plan for growth during an extended period of media disruption.
I have found the experience exhausting, sometimes frustrating but ultimately rewarding. Is it an approach I would recommend to others? Yes! Although disruptive in its own right, it can provide a fantastic opportunity to press the reset button, take advantage of new markets, invest in people and drive efficiencies. I feel proud to have led such a transformational project for the FT, one that will continue to bring benefits for years to come.